Researchers with the Lucy Family Institute for Data & Society at the University of Notre Dame have released a policy brief warning that artificial intelligence systems systematically distort historical information, posing a serious threat to democratic discourse and participation both nationally and internationally. The brief, developed in collaboration with the Keough School of Global Affairs at the University of Notre Dame, represents the first joint policy initiative between the two units.
The international research team, led by Nuno Moniz, Associate Research Professor, of the Lucy Family Institute, conducted the first systematic analysis of how AI systems present historical information. The interdisciplinary team includes Atalia Omer and Peter Cajka from the University of Notre Dame, Jasna Ćurković Nimac and Tomislav Anić from the Catholic University of Croatia, Silvana Mandolessi from the Catholic University of Leuven (Belgium), and Miguel Cardina from the Centre of Social Studies (University of Coimbra), combining historians, memory study scholars, philosophers and computer scientists. The team examined over 3,500 prompts across seven languages and four major conflicts: the US-Vietnam War, the Yugoslav Wars, the Portuguese Colonial Wars, and the Palestine/Israel conflict. The study revealed systematic problems including fabricated content, chronological distortions, bias through data prevalence, and language manipulation vulnerabilities.
The research reveals an urgent generational threat. Students today are already learning history through AI systems, with a 2025 survey finding that about a quarter of U.S. teens have used ChatGPT for schoolwork – double the share from 2023. The policy brief warns that without immediate action, future generations may be impacted in their understanding of the past and their ability to participate in democratic society.
The brief presents three urgent recommendations for U.S. federal and state governmental bodies. First, the researchers call for enhanced disclosure requirements when AI systems address historical topics, and for the Federal Department of Education to establish working groups creating professional development resources for educators. Second, the brief recommends that the Federal Trade Commission establish an independent review board to conduct yearly assessments of AI historical accuracy and for it to publish an annual Historical Accuracy Index. Third, the researchers urge the State Department to build capacity for identifying and responding to systematic historical misinformation campaigns, while calling on the National Science Foundation and Department of Defense to support research on tackling AI historical manipulation.
“The Lucy Family Institute is committed to advancing research that addresses society’s most pressing challenges at the intersection of data, AI, and society,” said Nitesh Chawla, the Frank M. Freimann Professor of Computer Science and Engineering, Founding Director of the Lucy Family Institute for Data & Society, and the Lucy Family Director for Data & AI Academic Strategy at the University of Notre Dame. “This groundbreaking work reveals how AI systems can fundamentally alter our shared understanding of history. As we continue to explore the societal implications of these technologies, we must act urgently to protect the foundations of civic education and democratic discourse that should be built on accurate historical knowledge.”
Moniz emphasized the narrow window for action. “The generational nature of this threat makes delayed action particularly costly,” said Moniz. “Many students that are currently at a university level are very likely to have incorporated AI-generated historical information into their foundational knowledge. Each additional year without intervention results in another student cohort whose historical understanding increasingly includes potentially misleading AI-generated content alongside traditional verified materials. We must address this immediately.”
The collaboration between the Lucy Family Institute and the Keough School of Global Affairs brings together expertise in data science and AI with policy implementation and global governance. “The Keough School’s scholarship on democracy, governance and human rights is enriched by the Lucy Family Institute’s technological expertise.” said Andrés Mejía Acosta, Kuster Family Associate Dean for Policy and Practice and Associate Professor of Political Economy of Development at the Keough School of Global Affairs. “Together, we can ensure that policymakers have the evidence-based insights they need to protect democratic institutions in an age of rapidly advancing AI. This is exactly the kind of interdisciplinary, policy-relevant work that Notre Dame is uniquely positioned to produce.”
The policy brief arrives as AI systems become increasingly integrated into educational settings nationwide. A 2024 survey found that 56% of educators expect AI use to increase in their schools, yet 58% still had no AI training as of late that year. This research demonstrates that the facility for narrative manipulation creates unprecedented vulnerabilities for cognitive warfare targeting American historical understanding and democratic institutions, but also its potential impact at an international level.
The full policy brief, “How Artificial Intelligence Can Change Our Past: Urgent Action on AI Distortion of Historical Information,” is available at https://doi.org/10.7274/30803867.
Contact:
Christine Grashorn, Program Director, Engagement and Strategic Storytelling
Lucy Family Institute for Data & Society / University of Notre Dame
cgrashor@nd.edu / 574.631.4856
lucyinstitute.nd.edu / @lucy_institute